In the US, federal agencies and recipients of federal funding are subject to two overlapping legal frameworks: Section 508 of the Rehabilitation Act (29 U.S.C. § 794d) — Requires federal agencies to make their electronic and information technology accessible to people with disabilities. The 2017 refresh of Section 508 aligns the technical requirements with WCAG 2.0 Level AA, which for PDF documents means compliance with PDF/UA (ISO 14289). Americans with Disabilities Act (ADA) — The DOJ's Title II rule (finalised April 2024) explicitly requires state and local government websites and mobile apps to meet WCAG 2.1 Level AA, covering downloadable PDFs. The rule became effective June 2024, with compliance dates of April 2026 (large agencies) and April 2027 (smaller agencies). EU: Web Accessibility Directive (2016/2102) — Requires public sector bodies in EU member states to meet EN 301 549, which references PDF/UA for PDF documents. The European Accessibility Act extends similar requirements to the private sector from June 2025. Key enforcement actions: the DOJ's ADA.gov publishes settlements with government entities over inaccessible digital content — PDF inaccessibility is a recurring finding. Agencies should validate government PDFs with the PAC (PDF Accessibility Checker) before publishing.

The Freedom of Information Act (FOIA, 5 U.S.C. § 552) requires federal agencies to disclose requested records upon request, with exemptions for certain categories (national security, privacy, law enforcement). When an agency discloses a document with exempted information, the exempted portions must be redacted from the released version. For PDF redaction under FOIA: True redaction is mandatory — Applying a black rectangle overlay does not satisfy FOIA requirements because the underlying text remains in the PDF content stream and is recoverable. The NSA's own Redacting with Confidence guidance (widely referenced by other agencies) explicitly requires content removal, not visual overlay. Government attorneys should use Adobe Acrobat Pro's Redact tool (which removes content) not the commenting tools (which create visual overlays only). Metadata must be scrubbed — PDF metadata may contain the document author, internal system identifiers, draft version information, or original filename — all potentially exempt information that must be removed before release. Use Adobe Acrobat's Sanitize Document tool (which removes metadata, JavaScript, attachments, and hidden data) as the final step before FOIA release. Verification workflow — After redacting, copy all text (Ctrl+A) and paste into a text editor to confirm exempt text is not present. Run ExifTool to check metadata. Test with the pdftotext command-line tool for scanned PDFs that have been OCR'd. Notable FOIA redaction failures include the 2019 Manafort filing (federal court) and a 2006 NSA surveillance disclosure — both cases where visual overlays were used instead of true redaction.

Government records have retention requirements measured in years to decades — in some cases permanently. Standard PDFs that reference external fonts, use proprietary features, or rely on encryption may not render correctly on future systems. PDF/A (ISO 19005) was designed specifically for long-term archiving and is widely specified for government records: - The US National Archives and Records Administration (NARA) specifies PDF/A-1 or PDF/A-2 as the preferred format for electronic records transferred to the National Archives. - The EU's IDABC/ISA² programme has promoted PDF/A across EU government document management. - Many state and local government records management systems require PDF/A for permanent or long-term retention categories. PDF/A requirements by profile: - **PDF/A-1** (ISO 19005-1): Strictest — fonts embedded, no encryption, no JavaScript, no external references, no transparency - **PDF/A-2** (ISO 19005-2): Adds support for transparency, JPEG2000 compression, digital signatures, and optional content - **PDF/A-3** (ISO 19005-3): Adds arbitrary embedded file attachments — used for e-invoicing (ZUGFeRD/Factur-X) and structured data embedding Validate government archive PDFs with veraPDF (open source, command-line, widely used in government document management systems).

Many government agencies publish data in PDF format that would be more useful as structured open data. The Open Data Handbook rates PDF as a poor format for open data because it requires specialised tools to extract the underlying data, and extraction quality degrades significantly with complex table structures. Where PDF publication is unavoidable, agencies can improve machine-readability by: - Generating PDFs from structured source data (not scanned or printed documents) so a text layer is present - Using tagged PDFs with properly marked tables (header cells, scope attributes) that can be extracted by table-aware PDF parsers - Publishing the underlying data simultaneously in CSV, JSON, or XML alongside the PDF presentation - Embedding structured data using PDF/A-3's attachment capability for financial and statistical reports The UK's GOV.UK content design guidance explicitly discourages publishing PDFs when HTML would serve the content better — PDFs are hard to update, harder to make accessible, and exclude users on mobile. This is a useful benchmark for any government web publisher.